I'm a coterm in Computer Science at Stanford, which is a graduate degree. I did my undergrad here as well in Symbolic Systems.
I recently started doing the 80,000 hours career planning course.
I'm most compelled by suffering-based ethics, though I still find negative utilitarianism unsatisfactory on a number of edge cases. This makes me less worried about X-risks and by extension less long-termist than seems to be the norm.
My shortlist of cause areas is the following:
though I remain uncertain, especially about the following things:
Have you done anything with the dataset of 990-PFs? I'd be interested in helping you with the data analysis if you want.
Do you think there's a number you would accept for how many people treated with psychotherapy would be "worth" the death of one child?
How did you measure efficacy?
How many AI Safety researchers would be enough? 80k emphasizes the fact that there are only 300 people working on this full-time, meaning that the problem is extremely neglected. How many people would have to be working on this problem for it to no longer be considered neglected?